Non-small cell lung cancer (NSCLC) is the leading cause of cancer morality in men and women in the United States, and the overall long-term survial is less than 15%. Pathologic stage I makes up 25-35% of NSCLC cases and has a good prognosis. However, cancer relapse and death rate in this subset is 35 to 50% by 5 years. Chemotherapy is beneficial for the treatment of several localized solid tumors after resection and may prove to be useful in the treatment of patients with stage I NSCLC. Thr purpose of this project is to define tissue and serum tumor markers in patients with stage I NSCLC which predict for early cancer recurrence. Pathologic stage I NSCLC was chosen for study to eliminate the significant influence of positive lymph nodes and distant metastases on survival. Immunohistochemical staining will identify potential tissue tumor markers and radioimmunoassay (RIA) or enzyme-link immunosorbent assay (ELISA) will identify potential serum tumor markers. Specific aim #1 will examine a set of twelve tissue tumor markers in a retrospective cohort of 275 stage I NSCLC patients. Markers are categorized by hypothetical method of action: molecular genetic markers (Kras, erbB-1, erbB-2, rb, p53, bcl-2), markers of metastatic propensity (angiogenesis factor viii), proliferation markers (K1-67) and markers of cellular differentiation (Blood group A, H/LeV/LeB, NCAM, CD44). Results will be used to develop a prediction rule for recurrence in stage I NSCLC using Cox proportional hazards regression analysis and an artificial neural network. Specific aim #2 will examine a set of eight serum tumor markers in a retrospective cohort of 250 patients with stage I to IV NSCLC. These markers are categorized as molecular genetic markers (anti-p53), markers of metastatic propensity (angiogenesis bFGF), somatamedins (growth factor IGF- 1) and markers of cellular differentiation (CEA, CA-125, CA 15-3, CYFRA21- 1, CD44). The purpose of this aim is to identify any correlations between titers of serum markers and tumor histology, stage or mass. One hundred patients in this cohort had a second serum collection after tumor resection. This subgroup of serum will allow analyses of titters before and after cyto-reduction. Significant correlates with tumor stage and mass will be evaluated in a prospective cohort of patients with stage I NSCLC. In specific aim #3, paraffin-embedded and fresh-frozen tumor tissue will be collected from a prospective cohort of 330 patients with stage I NSCLC to validate the prediction rule developed in specific aim #1. In these same patients, serial serum specimens will be collected for a minimum of 2.0 years after resection (specific aim #4). The significant markers identified in specific aim #2 will be analyzed in this cohort to describe correlations with tumor recurrence. Tissue and serum markers identified by the model can be used to select high risk patients for a prospective, multi-institutional chemotherapy trial for stage I NSCLC.